A bond left with nothing on the other end, on an atom at a surface that has lost the neighbour it was bonded to inside the crystal. Such bonds grab stray molecules and trap electrons, which is why ordinary surfaces are messy – and why layered crystals, whose sheets keep every bond inside themselves, split into faces that stay clean down to a .
Going deeper
Left: cutting a covalently bonded crystal leaves every surface atom with a bond pointing at nothing, which traps charge and reacts with whatever arrives. Right: a layered crystal splits between sheets whose bonds all stay inside them, so its faces have none left over; only edges and defects carry dangling bonds.
A bond with nothing on the other end
Inside a silicon crystal every atom shares four bonds with four neighbours. Cut the crystal and the atoms at the new surface lose some of those partners, and each lost partner leaves an orbital holding a single electron and pointing into empty space. These dangling bonds are costly, so the surface rearranges to pair them up where it can – the dimer rows of Si(100) and the 7×7 reconstruction of Si(111) are the result – and those that remain act as electronic states inside the and as chemically active sites that take up oxygen, water or anything else that arrives.
In a device the gap states are what matter. They trap and release charge, scatter carriers and pin the at the surface, so a moves the charge in the traps instead of the charge in the channel. Much of the history of the silicon is the history of getting rid of them: a thermally grown oxide, annealed in hydrogen, ties up nearly all of them, and that clean silicon–oxide interface is what made the MOSFET practical.
Surfaces with nothing cut
In a layered crystal the strong bonds all lie within a layer, and the layers are held together only by . Splitting it between layers cuts no bonds, so the exposed face – the basal plane – has none left dangling. That is why a single layer can exist on its own without its surface reconstructing, why graphene and hBN are atomically flat and chemically quite inert, and why hBN makes such a good : on it, graphene’s rose about tenfold compared with silicon dioxide. It is also why two different layers can be stacked in any combination, because with no bonds to join, their atomic spacings do not have to match.
The edges are the exception. A ’s edge, a or a missing atom does cut bonds, and that is where oxidise, catalyse reactions and host defect states. The activity of MoS2 comes from its edges rather than its basal plane, and a missing sulfur atom leaves molybdenum orbitals that trap electrons and give oxidation a place to start.
The price of an inert surface
A surface with nothing to bond to is hard to build on. grows oxide films by alternating two precursors that each react with the surface, and on a clean basal plane there is almost nothing to react with: growth starts only at steps, defects and contamination, and leaves islands and pinholes rather than a closed film. Seed layers – a thin evaporated metal oxidised in air, a molecular layer, a pre-treatment with ozone or plasma – restore nucleation, but each adds thickness, charge or damage to the interface it was meant to keep clean, which is why putting a thin on a 2D channel remains one of the main integration problems.
The alternatives avoid the chemistry altogether. Layered such as hBN, and oxide films grown elsewhere and laminated on top, form van der Waals interfaces with the channel and leave both surfaces intact; the difficulty moves to making them thin, uniform and . The same trade-off runs through contacts, and functionalisation: the inertness that keeps a 2D surface clean is what every process step has to work around.
For specialists
An unsaturated valence orbital at a surface, edge or defect where a covalent bond has been cut. In a bulk such as silicon these states lie in the gap, trap charge and pin the Fermi level unless passivated by hydrogen or a grown oxide. The basal plane of a layered crystal has none, which is what allows van der Waals stacking without lattice matching and interfaces with few traps – and also why atomic layer deposition nucleates poorly on it. Edges, grain boundaries and do carry them, and are where 2D materials are chemically active.